[Dates, names and location have been omitted to protect the guilty.]

I recently provided a client a set of analysis with estimates twice. The estimates were, of course, longer than they hoped but realistic and achievable. The third time they asked for someone else to provide them by the end of that very day. I think my estimates were unpopular and impossible to discredit due to the accompanying analysis. So instead of accepting that their anticipated estimate was incorrect they tried to use a different method to generate a more "favourable" number.

Each of these three pieces of work were sizable and mission critical.